392

What are the differences between #import and #include in Objective-C and are there times where you should use one over the other? Is one deprecated?

I was reading the following tutorial: http://www.otierney.net/objective-c.html#preamble and its paragraph about #import and #include seems to contradict itself or at least is unclear.

Ryan Guill
  • 12,814
  • 4
  • 34
  • 48

10 Answers10

364

There seems to be a lot of confusion regarding the preprocessor.

What the compiler does when it sees a #include that it replaces that line with the contents of the included files, no questions asked.

So if you have a file a.h with this contents:

typedef int my_number;

and a file b.c with this content:

#include "a.h"
#include "a.h"

the file b.c will be translated by the preprocessor before compilation to

typedef int my_number;
typedef int my_number;

which will result in a compiler error, since the type my_number is defined twice. Even though the definition is the same this is not allowed by the C language.

Since a header often is used in more than one place include guards usually are used in C. This looks like this:

 #ifndef _a_h_included_
 #define _a_h_included_

 typedef int my_number;

 #endif

The file b.c still would have the whole contents of the header in it twice after being preprocessed. But the second instance would be ignored since the macro _a_h_included_ would already have been defined.

This works really well, but has two drawbacks. First of all the include guards have to be written, and the macro name has to be different in every header. And secondly the compiler has still to look for the header file and read it as often as it is included.

Objective-C has the #import preprocessor instruction (it also can be used for C and C++ code with some compilers and options). This does almost the same as #include, but it also notes internally which file has already been included. The #import line is only replaced by the contents of the named file for the first time it is encountered. Every time after that it is just ignored.

Sven
  • 21,790
  • 4
  • 48
  • 70
  • 5
    This is the better answer than accepted one. @Guill, you should change the accepted answer. – Nguyen Minh Binh Jan 03 '13 at 09:18
  • 7
    After changing 4 `#include`s to `#import`s on a 7000 line template header file, there is a noticeable performance improvement in compilation and XCode intellisense responsiveness. (I don't think I'm imagining it) – bobobobo Aug 13 '13 at 18:42
349

The #import directive was added to Objective-C as an improved version of #include. Whether or not it's improved, however, is still a matter of debate. #import ensures that a file is only ever included once so that you never have a problem with recursive includes. However, most decent header files protect themselves against this anyway, so it's not really that much of a benefit.

Basically, it's up to you to decide which you want to use. I tend to #import headers for Objective-C things (like class definitions and such) and #include standard C stuff that I need. For example, one of my source files might look like this:

#import <Foundation/Foundation.h>

#include <asl.h>
#include <mach/mach.h>
Reto Koradi
  • 49,246
  • 7
  • 76
  • 116
Jason Coco
  • 76,627
  • 20
  • 180
  • 178
  • 65
    Even if header files contain include guards, there is still a performance hit during compilation if you use #include -- the compiler must open up each header file to notice the include guards. – Matt Dillard Jan 13 '09 at 16:41
  • 4
    a header guard is a preprocessor directive that ensures a header is only included once in a source file. – Jason Coco Jan 15 '09 at 16:14
  • 8
    I think #import is actually an addition by GCC, not by Objective-C. You can use it in non-ObjC languages as long as you compile with GCC (or Clang) – Dave DeLong Dec 01 '09 at 22:48
  • 34
    @dave - #import is an Objective-C addition to the preprocessor. GCC just supports it in C and C++ source files as well, although they officially suggest not using it in C or C++ in favor of portable, traditional header guards. All Objective-C preprocessors must include #import, however. – Jason Coco Dec 02 '09 at 04:01
  • 13
    A header guard is where you add to the top: `#ifndef myheader #define myheader` ... followed by header code... `#endif` – Tim Mar 12 '12 at 16:36
  • 2
    So the official recommendation is to avoid using the newer, simpler, better thing for the sake of maintaining compatibility with older things that could just as easily be updated to support the new keyword? That's nonsense. Somebody needs to bring in some new officials. – aroth Jan 24 '13 at 00:10
  • 3
    @aroth: It's not about newer vs. older. It's about standard vs. nonstandard. – Chuck Mar 07 '13 at 01:18
  • 2
    @Chuck - Perhaps. But a standard that doesn't evolve and that recommends dated, clunky, and obtuse ways of doing things is not very valuable, in my opinion. Especially when the capabilities of tools *do* evolve, and doubly so when "all Objective-C preprocessors must include #import" and when we're talking about code in the context of an Objective-C project. – aroth Mar 07 '13 at 03:32
  • 3
    On a decent compiler, the include guard won't cost at compile time. The header file needs to be included once if you want to use it. When it's included, the compiler will detect the header guard. and remember it. On the second #include for the same file, the compiler knows about the header guard and doesn't even look at the file. – gnasher729 Feb 23 '14 at 20:46
  • 7
    This is not standard vs. nonstandard; it's language vs. language, and one intent vs another. If you're using Objective-C **and** you intend to *include an Objective-C header*, use #import. If you're using C, C++ **or** are using Objective-C and just want to inline one file in another, use #include. – Steven Fisher Apr 16 '14 at 01:42
  • Another thing to note is that C++ has a different approach where you use `#pragma once` instead of manual header guards. So C++ still uses `#include` but also solves the problem `#import` is intended to solve. – uliwitness Aug 12 '15 at 17:31
  • @gnasher729 Right, this is called [Multiple-Include Optimisation](https://gcc.gnu.org/onlinedocs/cppinternals/Guard-Macros.html) and avoids a file to be unnecessarily reopened. – Bilow Dec 15 '16 at 20:47
  • 1
    This nice thing about `#import` that nobody here seems to mention, is that it's the only solution that doesn't require extra code. It isn't much code either way, but less code for the same thing is usually best. Also lowers the bar slightly for new devs who might not know about header guards or `#pragma once`. – William T Froggard May 15 '18 at 19:20
65

I agree with Jason.

I got caught out doing this:

#import <sys/time.h>  // to use gettimeofday() function
#import <time.h>      // to use time() function

For GNU gcc, it kept complaining that time() function was not defined.

So then I changed #import to #include and all went ok.

Reason:

You #import <sys/time.h>:
    <sys/time.h> includes only a part of <time.h> by using #defines

You #import <time.h>:
    No go. Even though only part of <time.h> was already included, as
    far as #import is concerned, that file is now already completely included.

Bottom line:

C/C++ headers traditionally includes parts of other include files.
So for C/C++ headers, use #include.
For objc/objc++ headers, use #import.

bdesham
  • 13,980
  • 10
  • 70
  • 116
user512705
  • 651
  • 5
  • 3
  • 2
    It seems that clang doesn't has this not-defined problem. – ooops Feb 23 '17 at 08:37
  • This is a really great example of why it's challenging to introduce modern functionality (#import) to existing systems in a non-intrusive way. Should the system libraries be updated to better handle #import ... probably not. If they were, it would be at the expense of many existing projects that knowingly or unknowingly rely on the existing behavior. Should language development stop introducing new, helpful, forward-looking features... no. So, it's never as simple as some of comments on the accepted answer suggest it could be. – rambo Aug 09 '20 at 18:53
24

#include works just like the C #include.

#import keeps track of which headers have already been included and is ignored if a header is imported more than once in a compilation unit. This makes it unnecessary to use header guards.

The bottom line is just use #import in Objective-C and don't worry if your headers wind up importing something more than once.

Ferruccio
  • 93,779
  • 37
  • 217
  • 294
13

I know this thread is old... but in "modern times".. there is a far superior "include strategy" via clang's @import modules - that is oft-overlooked..

Modules improve access to the API of software libraries by replacing the textual preprocessor inclusion model with a more robust, more efficient semantic model. From the user’s perspective, the code looks only slightly different, because one uses an import declaration rather than a #include preprocessor directive:

@import Darwin; // Like including all of /usr/include. @see /usr/include/module.map

or 

@import Foundation;  //  Like #import <Foundation/Foundation.h>
@import ObjectiveC;  //  Like #import <objc/runtime.h>

However, this module import behaves quite differently from the corresponding #include: when the compiler sees the module import above, it loads a binary representation of the module and makes its API available to the application directly. Preprocessor definitions that precede the import declaration have no impact on the API provided... because the module itself was compiled as a separate, standalone module. Additionally, any linker flags required to use the module will automatically be provided when the module is imported. This semantic import model addresses many of the problems of the preprocessor inclusion model.

To enable modules, pass the command-line flag -fmodules aka CLANG_ENABLE_MODULES in Xcode- at compile time. As mentioned above.. this strategy obviates ANY and ALL LDFLAGS. As in, you can REMOVE any "OTHER_LDFLAGS" settings, as well as any "Linking" phases..

enter image description here

I find compile / launch times to "feel" much snappier (or possibly, there's just less of a lag while "linking"?).. and also, provides a great opportunity to purge the now extraneous Project-Prefix.pch file, and corresponding build settings, GCC_INCREASE_PRECOMPILED_HEADER_SHARING, GCC_PRECOMPILE_PREFIX_HEADER, and GCC_PREFIX_HEADER, etc.

Also, while not well-documented… You can create module.maps for your own frameworks and include them in the same convenient fashion. You can take a look at my ObjC-Clang-Modules github repo for some examples of how to implement such miracles.

Alex Gray
  • 14,784
  • 6
  • 90
  • 114
4

If you are familiar with C++ and macros, then

#import "Class.h"

is similar to

{
#pragma once

#include "class.h"
}

which means that your Class will be loaded only once when your app runs.

LittleBobbyTables - Au Revoir
  • 30,184
  • 25
  • 98
  • 113
Evol Gate
  • 2,217
  • 3
  • 17
  • 36
  • Is this a supported use of #pragma once? I always thought the pragma needed to be *inside* the includ*ed* file to work. – uliwitness Aug 12 '15 at 17:33
  • @uliwitness You are correct. `#pragma once` is placed in the included file, not the file that performs the include. -1 for that. – herzbube Jun 22 '16 at 13:10
1

In may case I had a global variable in one of my .h files that was causing the problem, and I solved it by adding extern in front of it.

neowinston
  • 6,762
  • 9
  • 49
  • 80
0

#include it used to get "things" from another file to the one the #include is used in. Ex:

in file: main.cpp

#include "otherfile.h"

// some stuff here using otherfile.h objects,
// functions or classes declared inside

Header guard is used on the top of each header file (*.h) to prevent including the same file more then once (if it happens you will get compile errors).

in file: otherfile.h

#ifndef OTHERFILE
#define OTHERFILE

// declare functions, classes or objects here

#endif

even if you put #include "otherfile.h" n time in your code, this inside it will not be redeclared.

LPL
  • 16,007
  • 6
  • 43
  • 87
Celso Dantas
  • 9
  • 1
0

IF you #include a file two times in .h files than compiler will give error. But if you #import a file more than once compiler will ignore it.

Husmukh
  • 9
  • 2
  • 8
    `#include` the same file twice *does not* result in an error. – kennytm Jul 16 '10 at 09:45
  • 1
    To complement @KennyTM's comment, #include-ing the same file twice in the same header does not result in a compile error IF the usual header gards (#ifndef _FILE_NAME_H_ #define _FILE_NAME_H_ #end) are there. This is expected practice. Using #import the header guards aren't needed. – jbat100 Dec 17 '12 at 11:16
  • @jbat100: `#include` is simply a copy-and-paste mechanism. There is deliberate use of `#include` more than once without include guards, e.g. the "X macro". – kennytm Dec 18 '12 at 05:46
  • Including a file twice *may* result in errors depending on what you include. I've seen C code that used `#include` to implement a kind of templates. They did a `#define`, included a header, `#undef`d and redid the `#define`, included the same header a second time. This resulted in the code being parameterized, valid, and included twice, as the value of the define was different. So there are advantages to using `#include`, but if you're using a modern language like C++ or ObjC, you generally don't need this. – uliwitness Aug 12 '15 at 17:39
0

#include vs #import

History:

#include => #import => Precompiled Headers .pch => @import Module(ObjC); => import Module(Swift)

[Precompiled Headers .pch]
[@import Module(ObjC);]
[import Module(Swift)]

#include + guard == #import

#include guardWiki - macro guard, header guard or file guard prevents to double include a header by a preprocessor that can slow down a build time

#import disadvantage

Works with file scope that is why we have slow build time because a compiler must parse and compile as many times as many .h files were imported in project scope

[#import in .h or .m]

yoAlex5
  • 13,571
  • 5
  • 105
  • 98